The History Of The Automated Lung Radiograph

Published Date: 02 Nov 2017

SCREENING

Computer Aided Detection (CAD) is an emerging technique that involves the enhancement of the quality and the productivity of radiologists’ tasks by improving the accuracy and consistency in radiologic diagnosis, as well as reducing the image reading time [10]. In pulmonary medicine, the majority of CAD systems that are developed in the past and presented in the literature focus on the detection of lung nodules, with a view to automating lung cancer detection in chest radiographs and CT scan [17-22]. The current CAD systems can considerably enhance experienced radiologists’ performance and outweigh human limitations in identifying many diseases in earlier stages [11].

Problem Statement

Radiographic decision making is difficult, and in most cases is a subjective task. This is due to the potential overlap of abnormalities. Double reading is a proved medical practice for reducing the error rate in diagnostic imaging; thus, improving the detection accuracy.

A trained radiologist or a physician can be the most appropriate person for this role in cooperation with the responsible radiologist. However, small lung features (e.g., small nodules) are often missed in the interpretation [13], [14]. Erroneous interpretation of radiographs results from human subjectiveness, fatigue, and limited accuracy. Hence, the interest in developing an automated system for reliable and accurate detection at a high level of availability is crucial. Devising such a system for the detection of an infectious disease, such as TB, may overcome the inaccuracy associated with field interpretation; and consequently contributes to controlling the disease and providing better health care services to Canadians.

There are various difficulties associated with the detection of chest radiographic abnormalities that suggest the presence of pulmonary TB. The most apparent difficulty is the disconnected patterns of abnormality. We could, for example, have signs of nodules or cavities with a wide range of sizes- from few millimetres up to several centimetres- with large variation in density, and hence visibility on radiographs. Some abnormalities can be slightly denser than the surrounding lung tissue, while the densest ones are calcified. Furthermore, some features can be obscured by other organs, such as ribs, mediastinum, and structures below the diaphragm. Another difficulty is the possibility of overlapping the detection and diagnosis of TB with other lung diseases.

Global control of tuberculosis is hampered by slow, insensitive diagnostic methods, particularly for the detection of drug-resistant forms and in patients with human immunodeficiency virus infection. Early detection is essential to reduce the death rate and interrupt transmission [1].

Project Objective

The objective of this project is to consider and implement any image processing techniques that can aid in the automatic recognition of lung anomalies of all types, but must specifically include Tuberculosis. In order to achieve this, various sub-objectives need to be accomplished.

Segmentation of lung field.

Enhancement of lung field.

Recognition of lung anomalies.

Classification of lung anomalies.

Calibration and adjustment of various algorithms/values.

Segmentation of lung field

For the recognition of lung anomalies, only the lung segment of a chest radiograph needs to be considered. Thus an algorithm that segments the lung field from the rest of the chest radiograph needs to be implemented.

Enhancement of lung field

In order to make the recognition of lung anomalies easier and more accurate, the images obtained from the lung field segmentation needs to be enhanced.

Recognition of lung anomalies

Once the lung field has been segmented and enhanced, image processing algorithms/techniques need to be implemented that can recognise the various lung anomalies. Lung anomalies such as lung nodules and lungs with irregular seize/shape needs to be identified.

Classification of lung anomalies

The various recognized lung anomalies then needs to be classified, i.e. Cancer, Tuberculosis, Pneumoconiosis, etc.

Calibration and adjustment of various algorithms/values

Depending on the accuracy of the results obtained and the number of false positives, some of the algorithms and values (such as sensitivity) might need to be adjusted in order to increase the accuracy.

Once all the above sub-objectives are achieved, the overall project objective is accomplished.

Scope of the Project

The scope of the project involves all the sub-objectives stated above. There is a degree to which the scope extends and can be read below.

• Design and build an optical fibre vibration sensor. The design process will include research

• As mentioned above, the project objective is to demonstrate the operation of a simple fibre optic gyroscope. Hence, navigation type accuracy is not needed. This project serves as the basic building blocks for advancements at a later stage.

• Design and build a simple fibre optic gyroscope. The design process will include research into the various components needed, the design specifications and calculations. Once the design is finalised with all the components available, the fibre gyroscope will be constructed.

• Seeing as the project is centred on research and development, no electronic systems will be designed and built. Devices such as the detector and optical amplifiers (if needed) will be purchased. If components are available in the university laboratories, those components will be used.

• Open or closed loop designs of the optical fibre vibration sensor can be implemented as long as the principal is demonstrated.

• The optical fibre vibration sensor should be calibrated for single point sensing and again after multi-point sensing has been implemented.

• Once single point vibration sensing has been established, multi-point sensing should be implemented.

• Testing should be carried out to verify operability.

1.4.1 Assumptions.

1. Fibre optic cables are free from manufacturing defects.

2. Detectors are fully functional.

3. Most of the components are available in the optics laboratories.

Methodology Overview

Specifications and Requirements

The system and sub-systems refers to the certain requirements that are to be achieved by the end of the project. These are the project objectives and sub-objectives.

1.5.4 Design.

Taking the system and sub-system requirements into account, preliminary design idea are documented or researched. These ideas should then be refined to meet the specifications and a design is then selected. Once the design is selected, the planning for implementing the design should be done.

1.5.6 Construction.

When all the hardware and software are available, the construction process can begin. Here, the proposed design is implemented.

1.5.7 Testing.

The final product after construction is then tested. Specialised tests should be conducted to verify that the system designed, meets the requirements.

1.5.8 Evaluation.

At this stage, all the test results are analysed and evaluated.

1.5.9 Optimisation.

If it is found that certain improvements can be made to the design to yield more accurate results or performance, the optimisation phase feeds back to the design phase. At the design phase, the design is refined and improved. The process continues as normal from the design phase to the final product. If the system cannot be improved, that system will be the final product.

1.5.10 Final Product.

Only when the specifications have been met and the project objectives have been accomplished, can the product be termed the final product.

The advantage of the above methodology is that it has a well defined structured process. Furthermore, it has a logical flow.

A disadvantage however is that for optimisations, the entire design process has to be reviewed. This is very time consuming.

Deliverables

1. A full written report on the project (mini- dissertation).

2. Two seminars where the phases and project details will be presented.

3. A small working model of the optical fibre vibration sensor.

ECSA Outcomes

The following ECSA outcomes are tested in the module PJE:

1. Engineering problem solving:Learner must demonstrate competence to identify, assess,

formulate and solve open-ended engineering problems creatively and innovatively.

2. Application of fundamental and scientific knowledge:Learner must demonstrate

competence to apply knowledge of mathematics, basic science and engineering science from

first principles to solve engineering problems.

3. Engineering design and synthesis:Learner must demonstrate competence to perform

creative, procedural or non-procedural design and synthesis of components, systems,

engineering works, products or processes.

4. Investigations, experiments and data analysis:Learner must demonstrate competence to

design and conduct investigations and experiments.

5. Engineering methods, skills and tools, including Information Technology:Learner must

demonstrate competence to use appropriate engineering methods, skills and tools, including

those based on information technology.

6. Professional and technical communication:Learner must demonstrate competence to

communicate effectively, both orally and in writing, with engineering audiences and the

community at large.

Suggested length: 1 page.

Overview of the Document

There are eight chapters in this document. A brief overview of these chapters can be read below.

Chapter 2: Requirements Analysis.

The problem gets discussed in more detail in which the issues, constraints and requirements are discussed. It includes all the technical, legal, safety social aspects as well as various considerations. Furthermore, the requirements specification is presented in this chapter. This states what must be done for the project objective to be accomplished and how it is to be done.

Chapter 3: Literature Study.

The current state of optical fibre vibration sensor is discussed in this chapter. A full literature study on the theory, tools, similar works and software applicable to optical fibre vibration sensors are presented in this chapter.

Chapter 4: Design.

This chapter deals with the overall design process. Various designs are evaluated and a final design is selected. A detailed design of the proposed solution is presented in this chapter.

Chapter 5: Experimental design.

Test and evaluation strategies are defined in this chapter. The overall experimental procedure is documented in this chapter. This includes the experimental design, equipment used, the setup and the expected results and analysis.

Chapter 6: Implementation Overview.

All the issues, as well as cost, encountered in the implementation of the design are discussed in this chapter. The issues include integration issues, construction issues etc. Solutions to these problems are also presented.

Chapter 7: Results and Analysis.

This chapter deals with the analysis of results obtained during testing.

Chapter 8: Conclusion.

This is the last chapter and it gives an overview of the entire project from beginning to end. Discussions on various aspects of the project are discussed in this chapter. It sums up the achievement of the project objective.

Conclusion